Cyclic adsorption processes are well known and are typically used to separate a more adsorbable component from a less adsorbable component. The typical cyclic adsorption process employs a selective adsorbent to remove at least one component of a gas mixture, employing four basic process steps: (1) adsorption, (2) depressurization, (3) purge and, (4) pressurization. The feed fluid, usually a gas, containing the more readily adsorbable component and a less readily adsorbable component is passed through at least one adsorbent bed capable of selectively adsorbing the more readily adsorbable component at a predetermined (upper) adsorption pressure. The stream exiting the bed at this upper pressure is now concentrated in the less readily adsorbable component, and is removed as product. When the bed becomes saturated with the readily adsorbable component, the bed is thereafter depressurized to a lower desorption pressure for the desorption of the readily adsorbable component, with this component discharged from the system. Such processes are generally used to separate gases such as oxygen or nitrogen from air; hydrocarbons and/or water vapor from feed air gases; hydrogen from carbon monoxide; carbon oxides from other gas mixtures; and the like.
Examples of suitable cyclic adsorption system include, but are not limited to, pressure swing adsorption (PSA), vacuum swing adsorption (VSA) or vacuum pressure swing adsorption (VPSA) processes which use a low pressure or a vacuum and a purge gas to regenerate the sorbent and temperature swing adsorption (TSA) processes which uses a thermal driving force such as a heated purge gas to desorb the impurities.
Traditionally, cyclic adsorption plants, such as VPSA plants, use positive displacement machines such as rotary lobe type blowers operating at fixed speeds to move gas through the process. These machines are robust and generally do not experience any significant operational problems as the pressures and flows change and reverse. However, these machines have low power efficiency and are typically only 60-65% efficient.
More recently, applicants and others have proposed the use of more efficient machines capable of meeting the rigorous requirements of rapid cyclic conditions in place of traditional rotary lobe type machines. For example, U.S. Pat. No. 7,785,405B2 discloses systems and processes for gas separation using high-speed permanent magnet variable-speed motors to accelerate and decelerate centrifugal compressors used in PSA or VPSA processes. The centrifugal compressors are driven by direct drive variable, high speed permanent magnet motors or direct drive variable, high speed induction motors and have efficiencies of approximately 85%. Such compressors can accelerate from low-speed to high-speed and decelerate from high-speed to low-speed at very rapid rates offering a magnitude improvement over the capabilities of conventional machines with conventional motor/gear box systems.
One challenge in using centrifugal compressors is that the compressor performance is very sensitive to rapid changes in pressure, such as the pressure changes that typically occur during cyclic adsorption processes. When the process cycle requires the compressor speed to decelerate due to falling pressure ratios, the control system or controller typically direct the variable frequency drive (VFD) to disable energy input to the motor allowing the drive train (motor rotor and compressor impeller) to “free-wheel” decelerate (coast) down to its minimum speed without consuming power. If the drive train reaches the minimum speed too quickly, such as before the completion of the falling pressurization equalization step, the VFD re-enables energy input to the motor thereby consuming unneeded power. Power as used herein refers to electrical power.
It has now been found that by properly operating the machine during deceleration one can avoid reaching the design minimum operating speed too quickly. By avoiding reaching minimum operating speed before the pressure ratio across the machine starts rising, the machine and can be operated at peak efficiencies. Thus, one objective of this invention is to match the deceleration rate of the compressor/drive train to the decreasing pressures ratio across the machine so that the centrifugal machine arrives at its minimum operating speed near the point required to begin acceleration/reacceleration and, preferably, at the point required to accelerate/reaccelerate and along the compressor's best efficiency line (as shown in FIG. 5). This eliminates the unnecessary power consumption used during machine idling time which occurs after deceleration as further described below.
One method to match the compressor deceleration speeds to the decreasing pressures ratio is through the use of sophisticated control systems that continuously measure, or monitor at frequent points, the machine speed and the pressure ratio. The rotation of the drive train of the centrifugal machine is then controlled using dynamic braking (energy is fed to a braking resistor) or regenerative braking (energy is fed back into the power grid or stored in a flywheel until needed) based upon instructions received from the control system to reduce the deceleration rate and arrive at its minimum operating speed near or at the point required to begin acceleration/reacceleration.
However, the preferred method to control the rotation of the drive train is through the use of aerodynamic braking which can be controlled by either increasing or decreasing the feed fluid mass flow to the compressor which in turn increases or decreases the amount of work done to the fluid by the compressor impeller. This is accomplished by the use of flow control valves, preferably at the suction inlet of the compressor, and matching the mass flow to the desired deceleration rate through the operation and movement of the valves. The valves are opened to release process fluid and thereby the mass flow or closed to reduce mass flow. In this way, the aerodynamic braking that occurs during free-wheel deceleration can be reduced by suction throttling one or more valves at fixed position. One or more control valves can be employed depending on the system and number of adsorber vessels employed. This approach is economical and simple to commercially implement.